We present national prevalence rates for work organization characteristics for the current main job held by sample adults or for the most recent job held by sample adults not working at the time of interview, but who worked at some time in the previous 12 months. Employment was defined as working for pay at a job or business or working, but not for pay, at a family-owned business or farm. To ensure that respondents answered about the job of interest, questions, and question sets often used a lead-in similar to the following: “The next few questions refer to [fill: your job as a (JOB DESCRIPTION) with (EMPLOYER NAME)/your current, MAIN job/the job you held [most recently].” We then classified current/recent workers by demographic (sex, age, race/ethnicity, marital status, education, and class of worker) and geographic characteristics (place of residence and region). Geographic classification was based on the location of the respondent’s home as within or outside a metropolitan statistical area (MSA). Analysis by educational status was limited to workers aged 25 years and older. Industry and occupation categories were created by NCHS based on the North American Industry Classification System (NAICS) and the Standard Occupational Classification System (SOC) codes.

We looked at the following work organization, or structure of work characteristics: long work hours, nonstandard work arrangements, temporary position, and alternative shifts. Our calculation of prevalence rates for long work hours were restricted to currently employed adults working one job for two reasons: first, “hours worked per week” was asked only of currently employed adults; and second, this question did not distinguish between main jobs and secondary jobs, whereas all other work organization questions referred to a main job. Therefore, to be consistent with the other outcomes which focused on main jobs, we excluded employed adults working two or more jobs from our analyses of long work hours. We used two definitions for long work hours. First we looked at currently employed adults who reported working, on average, 48 hr or more per week. Then we looked at currently employed adults who met a more extreme definition of long work hours, 60 hr or more per week.

For all other work organization characteristics, currently employed adults holding more than one job were included but were asked to consider only their main job when answering the questions. Non-standard work arrangements were measured by asking “Which of the following best describes your work arrangement?” The following responses were all considered to represent nonstandard work arrangements: (1) “you work/worked as an independent contractor, independent consultant, or freelance worker;” (2) “you are/were on-call and work/ worked only when called to work;” (3) “you are/were paid by a temporary agency;” (4) “you work/worked for a contractor who provides workers and services to others under contract;” and (5) “other [work arrangement].” Those indicating that they are/were a regular permanent employee were considered to have a standard work arrangement. Temporary position was measured by a positive response to the following question: “Some people are in temporary jobs that last only for a limited time or until completion of a project,—Is/Was your job temporary?” Alternative shift was defined as those answering the question “Which of the following best describes the hours you usually work/worked?” with responses of—”a regular evening shift,” “a regular night shift,” a rotating shift,” or “some other schedule?” Those who indicated working “a regular day shift” were the comparison group.

Tetrachoric correlations between work organization factors are provided in Appendix A.

All analyses were conducted using SAS-callable SUDAAN software version 10.0 [RTI, 2008] to account for the complex sampling design of the NHIS. To represent the U.S. civilian, non-institutionalized population age 18 years and over, and to estimate the total number of employed US civilian workers represented by each individual in the sample, all estimates were weighted using the NHIS sample adult record weight. Point estimates with a relative standard error (RSE) greater than 30% but less than or equal to 50% are noted in the text and indicated with an in the tables as they do not meet the NCHS standards of reliability/precision. Estimates with a RSE greater than 50% or based on cell sizes less than 10 cases are not shown.

In order to assess patterns of prevalence for work organization characteristics among workers by industry and occupation group, we ranked groups from highest to lowest unadjusted prevalence rate. Note that these rankings do not account for whether or not the differences between estimates were statistically significant. However, we did calculate significance tests that tested for statistically significant differences between the industry and occupation groups with the highest prevalence rates for work organization characteristics, and the prevalence rate of these characteristics for all current/recent workers combined. These significance tests were adjusted such that the estimated standard error of the difference between prevalence rates for industry and occupation groups and all current/ recent workers accounted for non-independence of industry and occupation groups and all current/recent workers by incorporating their covariance [a method used in Cohen and Makuc, 2008]. Differences that were statistically significant (P < 0.05) for select variables are noted in the text.

When examining the prevalence rate of work organization factors among various industry and occupation groups, we present unadjusted prevalence rates that may be useful for comparisons to unadjusted data from other sources (e.g., Occupational Information Network (O*NET)), and for identifying groups of workers with the higher burdens of exposure to target with preventive strategies. Some researchers may prefer to use adjusted prevalence rates for industry and occupation groups to make our estimates comparable to those of the Quality of Employment Surveys, which was adjusted for age, sex, and race [Karasek et al., 1988; Schwartz et al., 1988; Pieper et al., 1989; Reed et al., 1989; Alterman et al., 1994]. In Table I we present prevalence rates adjusted by age, sex, and race/ethnicity using the projected 2000 U.S. population as the standard population [Day, 1996]. Although we do not discuss individual adjusted prevalence rates in this article due to space limitations, we are making them available for researchers to use.

The overall prevalence rate of non-standard work arrangements was 18.7% (Table II). Higher prevalence rates for non-standard work arrangements were found for men (20.9%) compared with women (16.4%; P < 0.01); workers aged 65 or older (37.7%) compared with other age groups P < 0.01 for all pair-wise comparisons); Hispanic workers (21.7%) compared with other racial/ethnic groups (P < 0.05 for all pair-wise comparisons), excluding non-Hispanic other adults; and those having less than a high school diploma (25.3%) compared with those having more education (P < 0.01 for all pair-wise comparisons). The prevalence rate of non-standard work arrangements was also higher among those who were self-employed in their own business, professional practice, or farm (83.9%) compared to employees of private companies (11.8%; P < 0.01) and employees of Federal, state, or local governments (11.1%; P < 0.01).

Compared to all adults employed at some time in the past 12 months, higher industry prevalence rates for nonstandard work arrangements (see Table III) were identified for workers in Construction (44.1%; P < 0.01), followed by Agriculture, Forestry, and Fishing (42.8%; P < 0.01) and Administrative Support and Waste Management and Remediation Service industries (38.8%; P < 0.01). The prevalence rate of non-standard work arrangements by occupation was higher for those in Arts, Design, Entertainment, Sports, and Media occupations (47.4%; P < 0.01); followed by Construction and Extraction occupations (44.0%; P < 0.01); and Personal Care and Service occupations (38.2%; P < 0.01).

The overall prevalence rate of workers working in temporary positions was 7.2% (Table II). Higher prevalence rates were found among men (7.8%) compared with women (6.6%; P < 0.05); workers aged 18–29 (12.3%) compared with other age groups (P < 0.05 for all pair-wise comparisons); Hispanic adults (10.6%) compared with non-Hispanic white workers (6.2%; P < 0.01) and non-Hispanic Asian, Native Hawaiian, or other Pacific Islander workers (6.7%; P < 0.01); and those not having a high school diploma (10.4%) compared with those having more education (P < 0.01 for all pair-wise comparisons). Prevalence rates for temporary positions were also higher for those who were self-employed in their own business, professional practice, or farm (10.4%) compared to employees of private companies (6.5%; P < 0.01).

As shown in Table III, prevalence rates for temporary work positions were higher in Administrative and Support and Waste Management and Remediation Services (17.1%) compared to the prevalence rate for all adults employed at some time in the past 12 months (7.2%; P < 0.01). Higher prevalence rates were also found for workers in Arts, Entertainment, and Recreation (16.4%; P < 0.01); and Construction industries (15.1%; P < 0.01). Among occupational groups, workers in Farming, Fishing, and Forestry had a higher prevalence rate of temporary positions (25.9%; P < 0.01) compared to all adults employed at some time in the past 12 months. Similar findings emerged for Arts, Design, Entertainment, Sports, and Media (18.3%; P < 0.01); and Construction and Extraction occupations (17.4%; P < 0.01).

As shown in Table II, the overall prevalence rate for alternative shifts was 28.7%. Prevalence rates were higher for workers aged 18–29 (43.0%) compared with other age groups (P < 0.01 for all pair-wise comparisons); non-Hispanic blacks (34.5%) compared to non-Hispanic whites 28.1%; P < 0.01), non-Hispanic Asian, Native Hawaiian, and other Pacific Islander workers (26.2%; P < 0.01), and Hispanic workers (27.7%; P < 0.01). Workers with a Bachelor’s degree and higher (19.0%) had a lower prevalence rate of alternative shifts compared to workers with less education (P < 0.01 for all pair-wise comparisons). Among classes of workers, persons working without pay in a family owned business or farm reported a higher prevalence rate of working an alternative shift (55.4%) compared with other classes of workers (P < 0.05 for all pair-wise comparisons).

Compared to all adults employed at some time in the past 12 months, workers in the Accommodation and Food Services (62.3%; P < 0.01); Arts, Entertainment, and Recreation (49.6%; P < 0.01); and Retail Trade industries (48.8%; P < 0.01) had a higher prevalence rate of alternative shifts (see Table III). Food Preparation and Serving occupations had a higher prevalence rate (63.1%; P < 0.01) of alternative shift work; followed by protective service occupations (54.3%; P < 0.01); and sales and related occupations (45.9%; P < 0.01).

In 2002 the NIOSH Organization of Work Group published a report [Caruso et al., 2006] that discussed changes in the nature of work organization, and recommended improved measurement of working hours, and a focus on populations more likely to work long hours. In the current study, we examined workers working 48 hr or more, and working 60 hr or more separately because researchers have used many different definitions of long work hours, and because the health implications of each may be different. While the overall prevalence rate of working 48 hr or more was 18.7% and the prevalence rate of working 60 hr or more was 7.2%, the demographic, geographic, industry, and occupation groups with higher prevalence rates of exposure were nearly identical at each of the two cut points. However, we may have underestimated the prevalence rate for long work hours by restricting analyses to currently employed adults working only one job.

Our findings are similar to those of Grosch et al. [2006] in an analysis of 2002 data (n = 1,744) from the Quality of Worklife Survey (QWL) with regard to characteristics of those working long hours (male, white, married, college educated, and self-employed). However, the age group with a higher prevalence rate of workers working long hours in our study was ages 30–44; while younger workers, ages 18–34, were most likely to report long work hours in the QWL. Again, this difference may be explained by younger workers having multiple jobs. The QWL question included all jobs, while ours was restricted to one job. Grosch et al. [2006] also found a low prevalence rate of long working hours among Farming, Fishing, and Forestry occupations which is contrary to our findings of high prevalence rates of long work hours among this group.

For the purposes of this overview article, non-standard work was used as a broad term to include working as an independent contractor; independent consultant or freelance work; working on-call or only when called to work; working for a temporary agency; and working for a contractor who provides workers and services under contract. Non-standard work arrangements expose workers to “precarious work” often characterized by reduced wages, status, security, and benefits, such as pension, insurance, and sick leave [Benach and Muntaner, 2007; Hadden et al., 2007]. On the other hand, non-standard work arrangements may benefit some workers by allowing them to control their schedules; and may benefit employers by providing an opportunity to screen workers prior to hiring them, and to cut labor costs during slack times.

According to the Current Population Survey (CPS), in 2005 non-standard work arrangements represented 10.7% of total employment broken down as follows: independent contractors (7.4%); on-call workers (1.8%); temporary agency workers (0.9%); and workers provided by contract firms (0.6%). The CPS, conducted by the U.S. Census Bureau for the Bureau of Labor Statistics (BLS), is a nationwide survey of 60,000 households that obtains information on employment, unemployment, earnings, and demographics of the civilian non-institutionalized population. In the current study the employment breakdown is similar: independent contractors (9.7%); on call workers (2.7%); temporary agency workers (1.0%); and working for a contractor (1.7%). Although the study methods differ somewhat, the overall prevalence rate of non-standard work arrangements in our study was 18.7%, suggesting that exposure to non-standard work arrangements may be increasing.

Temporary workers are often referred to as contingent workers who do not expect their jobs to last, or who report that their jobs are temporary. Temporary work may play an important role in the U.S. economy as a bridge to permanent employment for those who are out of work, or changing jobs. Temporary employment can range from a day or less, to several years. Temporary workers have grown in importance as firms have relied on them to meet their changing labor needs [Luo et al., 2010].

In our study, the overall prevalence rate of workers working in temporary positions was 7.2%. This is higher than the estimate of 1.8–4.1% from the 2005 CPS [BLS, 2005]. Since our definition of temporary workers closely follows the CPS definition of contingent work, it appears that this may be increasing. Similar to what we found in our study, data from the CPS showed that contingent workers were twice as likely to be under 25 years old, were more likely to be Hispanic, have less than a high school diploma, and more likely to be self-employed compared to non-contingent workers. However, the CPS study found contingent work to be more common among women than men, which is the opposite of what we found. Some of the industry and occupation groups found to have a higher prevalence rate of contingent/temporary work were similar between the 2005 CPS study and the present (2010 NHIS–OHS) study: education services and construction and extraction occupations [BLS, 2005].

The 2010 NHIS–OHS question on shift work captures data on evening and night shifts, along with rotating shifts, so that the prevalence rate of exposure to these alternative shifts could be determined and associations between this exposure and health outcomes may be examined in future studies. Our study found an overall prevalence rate of alternative shift work of 28.7%. In comparison, data from 2004 collected by BLS [McMenamin, 2007] indicated that 17.7% of workers worked alternate shifts that fell at least partially outside of the daytime shift range between 6 am and 6 pm. More than half of the full-time workers who worked an alternate shift in May 2004 reported doing so because it was “the nature of the job.” The 2004 results are similar to findings from the 2010 NHIS–OHS, with a higher proportion of workers working alternative shifts in Accommodation and Food Services industries, (52.7% in BLS compared with 62.3% in NHIS–OHS). Other industry groups with large proportions of employees who reported working alternate shifts in the BLS and the 2010 NHIS–OHS included Arts Entertainment and Recreation; Mining; and Transportation and Warehousing. Regarding occupation groups both the BLS report and the 2010 NHIS-OHS found that workers in Service occupations, especially Protective Service and Food Preparation and Serving occupations, were most likely to work alternate shifts. In general, the prevalence rate of alternative shift work for each group was higher in the NHIS–OHS than in the BLS sample. Prevalence rate differences may be due in part to the six-year time difference between the two surveys and to the increased use of flexible or alternative work schedules in recent years.

This study is subject to limitations often found in cross-sectional interview surveys. Because the focus was on a current or most recent job, data on changes in work organization characteristics over time are not available. Although the population-based sample design of the NHIS allowed us to make nationally representative estimates for many variables, small numbers of respondents with specific work organization characteristics, especially within certain demographic, industry, and occupation subgroups, made some estimates unstable (e.g., characteristics for Management of Companies and Enterprises), and several exposures for working long hours without pay in a family owned business or farm.

There are limitations associated with the industry and occupation groups used in these analyses. Broad industry and occupation categories may lump together workers who likely have substantially different workplace exposures. Conversely, small sample sizes within some industry and occupation groups result in wide confidence intervals that may result in underestimation or overestimation of exposure. Finally, the economic climate and high unemployment rates in the United States during 2010 should also be considered when interpreting these findings as such conditions could have potentially influenced the NHIS–OHS estimates.

Limitations aside, our study has a number of strengths. National prevalence estimates of work organization characteristics can be used to help occupational health professionals and employers to identify emerging occupational safety and health risks. In addition, the publication of nationally representative unadjusted and adjusted prevalence rates for work organization factors for multiple industries and occupations will allow researchers to use these data to impute work organization into their data by occupation or industry title [e.g., Schwartz et al., 1988; Alterman et al., 2008; Cifuentes et al., 2010] when exposure data is lacking.

Researchers have recently published several articles on work organization factors using this public use dataset, the first focused on shift work and short sleep duration [CDC, 2012], and the second examined gender differences in the effect of weekly working hours on occupational injury [Wirtz et al., 2012]. In the future we plan to examine the association of these work organization characteristics with several health outcomes. Understanding these associations may point to additional opportunities for prevention.